INDEX

Explanations

terms related to personal relationships and emotional experiences

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

Follows a personal pronoun and "had"

np_acts-logits-general · gemini-2.0-flash

pronouns followed by past tense verbs

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 furent

-0.86

 firent

-0.82

 went

-0.82

 took

-0.78

 furono

-0.77

 Went

-0.76

 became

-0.76

fue

-0.75

went

-0.74

 Became

-0.73

POSITIVE LOGITS

 hadn

1.10

 would

1.02

would

0.86

 habían

0.84

Would

0.81

 Would

0.79

had

0.79

 WOULD

0.78

 había

0.73

had

0.73

Activations Density 0.750%

terms related to personal relationships and emotional experiences

Follows a personal pronoun and "had"

pronouns followed by past tense verbs

No Comments

No Known Activations