INDEX

Explanations

expressions of confidence

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

states of confidence

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.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

 ujednoznacz

-0.54

 verrez

-0.53

deals

-0.52

 MainAxisSize

-0.52

jeras

-0.51

 resourceCulture

-0.50

 auroit

-0.48

ugl

-0.47

 vôtre

-0.46

 Stair

-0.46

POSITIVE LOGITS

 confident

2.06

confident

1.83

 Confidence

1.60

 confidence

1.58

Confidence

1.52

 confidently

1.50

confidence

1.47

自信

1.13

 confianza

1.10

 уверен

1.02

Activations Density 0.008%

expressions of confidence

states of confidence

No Comments

No Known Activations