-
Notifications
You must be signed in to change notification settings - Fork 677
All issues
Issue creation is restricted in this repository
Issues
is:issue state:open
is:issue state:open
Search results
[Bug Report] hook_attn_scores masks with finfo.min instead of -inf, so isinf-based mask detection silently finds nothing
bugSomething isn't workingSomething isn't workingcomplexity-simpleSimple issues, which may be good for beginnersSimple issues, which may be good for beginnersgood first issueGood for newcomersGood for newcomershelp wantedExtra attention is neededExtra attention is neededTransformerBridgeBug specific to the new TransformerBridge systemBug specific to the new TransformerBridge systemStatus: Open.#1692 In TransformerLensOrg/TransformerLens;[Bug Report] n_params_total does not mirror HookedTransformer — split and fused QKV are both counted
bugSomething isn't workingSomething isn't workingcomplexity-simpleSimple issues, which may be good for beginnersSimple issues, which may be good for beginnersgood first issueGood for newcomersGood for newcomershelp wantedExtra attention is neededExtra attention is neededTransformerBridgeBug specific to the new TransformerBridge systemBug specific to the new TransformerBridge systemStatus: Open.[Bug Report] Assigning cfg.use_attn_result / use_hook_mlp_in / use_split_qkv_input directly is silently ignored
bugSomething isn't workingSomething isn't workingcomplexity-highVery complicated changes for people to address who are quite familiar with the codeVery complicated changes for people to address who are quite familiar with the codehelp wantedExtra attention is neededExtra attention is neededTransformerBridgeBug specific to the new TransformerBridge systemBug specific to the new TransformerBridge systemStatus: Open.[Bug Report] Hooks on gated hook points are accepted and silently never fire (check_hooks_to_add is a no-op)
bugSomething isn't workingSomething isn't workingcomplexity-moderateModerately complicated issues for people who have intermediate experience with the codeModerately complicated issues for people who have intermediate experience with the codegood first issueGood for newcomersGood for newcomershelp wantedExtra attention is neededExtra attention is neededTransformerBridgeBug specific to the new TransformerBridge systemBug specific to the new TransformerBridge systemStatus: Open.[Proposal] Backward Lens: vocabulary readout of multilayer perceptron gradient factors
complexity-highVery complicated changes for people to address who are quite familiar with the codeVery complicated changes for people to address who are quite familiar with the codeenhancementNew feature or requestNew feature or requestTransformerBridgeBug specific to the new TransformerBridge systemBug specific to the new TransformerBridge systemStatus: Open.#1686 In TransformerLensOrg/TransformerLens;[Proposal] Add a bridge-vs-HF numerical parity test for BERT
complexity-moderateModerately complicated issues for people who have intermediate experience with the codeModerately complicated issues for people who have intermediate experience with the codehelp wantedExtra attention is neededExtra attention is neededtestingA task that needs to be completed in order to improve the current test coverage.A task that needs to be completed in order to improve the current test coverage.TransformerBridgeBug specific to the new TransformerBridge systemBug specific to the new TransformerBridge systemStatus: Open.[Bug Report] TransformerBridge.load_state_dict(assign=True) desyncs split QKV/gate-up components from the combined weight they view into
bugSomething isn't workingSomething isn't workingcomplexity-moderateModerately complicated issues for people who have intermediate experience with the codeModerately complicated issues for people who have intermediate experience with the codeTransformerBridgeBug specific to the new TransformerBridge systemBug specific to the new TransformerBridge systemStatus: Open.#1637 In TransformerLensOrg/TransformerLens;[Bug Report] Native init drifts from
HookedTransformer: unresolvedinitializer_rangesentinel and ignored xavier/kaiming gaincomplexity-moderateModerately complicated issues for people who have intermediate experience with the codeModerately complicated issues for people who have intermediate experience with the codegood first issueGood for newcomersGood for newcomershelp wantedExtra attention is neededExtra attention is neededTransformerBridgeBug specific to the new TransformerBridge systemBug specific to the new TransformerBridge systemStatus: Open.[Bug Report] [macOS-arm64] Cached eager attention NaNs in transformers v5 — blocks bridge KV-cache generation
discussionNo action needed yetNo action needed yetseen_by_maintainersConfirms that a maintainer is aware of this card.Confirms that a maintainer is aware of this card.toolingAnything pertaining to outside tools used within the codebaseAnything pertaining to outside tools used within the codebasewontfixThis will not be worked onThis will not be worked onStatus: Open.#1322 In TransformerLensOrg/TransformerLens;[Question] How do I add a custom generative video transformer into TransformerLens?
complexity-highVery complicated changes for people to address who are quite familiar with the codeVery complicated changes for people to address who are quite familiar with the codeStatus: Open.#869 In TransformerLensOrg/TransformerLens;[Proposal] Add support for TracrBench
complexity-highVery complicated changes for people to address who are quite familiar with the codeVery complicated changes for people to address who are quite familiar with the codenew-architectureThis card involves adding a new architecture .This card involves adding a new architecture .Status: Open.#704 In TransformerLensOrg/TransformerLens;[Bug Report] Demo for Tracr to TransformerLens is kind of broken
bugSomething isn't workingSomething isn't workingdemoCreating a demo or tutorialCreating a demo or tutorialStatus: Open.#481 In TransformerLensOrg/TransformerLens;