INDEX

Explanations

conjunctions and transition words indicating causality and conclusion

oai_token-act-pair · gpt-4o-mini Triggered by @bot

causal conjunctions followed by outcome

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ereich

-0.48

 coû

-0.44

 nég

-0.41

 lobos

-0.38

ριά

-0.37

Professional

-0.36

████

-0.36

ībā

-0.35

StateChange

-0.35

scary

-0.35

POSITIVE LOGITS



0.82

tvguidetime

0.82

期刊论文

0.70

ConstraintMaker

0.64

DebuggerNonUser

0.64

 surla

0.62

 المعيارى

0.60

MigrationBuilder

0.59

참고

0.59

Życiorys

0.58

Activations Density 0.028%

conjunctions and transition words indicating causality and conclusion

causal conjunctions followed by outcome

No Comments

No Known Activations