INDEX

Explanations

verbs and nouns related to human behaviors and characteristics

oai_token-act-pair · gpt-3.5-turbo

mentions of behavioral patterns or inclinations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

mberg

-0.79

estamp

-0.78

aban

-0.77

gur

-0.75

loo

-0.75

riel

-0.75

han

-0.72

tein

-0.71

ÄŁ

-0.71

chn

-0.71

POSITIVE LOGITS

 tendency

1.04

 tendencies

0.95

 toward

0.92

 towards

0.90

 aversion

0.79

 shift

0.79

 impulse

0.75

 propensity

0.75

 urge

0.74

 revers

0.74

Activations Density 0.013%

verbs and nouns related to human behaviors and characteristics

mentions of behavioral patterns or inclinations

No Comments

No Known Activations