INDEX

Explanations

phrases indicating honesty or admission of truth

oai_token-act-pair · gpt-3.5-turbo

phrases that express honesty and admission

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

prus

-0.70

und

-0.66

ighth

-0.63

Fall

-0.62

world

-0.62

mb

-0.61

CONT

-0.61

USS

-0.60

andel

-0.59

åĮ

-0.59

POSITIVE LOGITS

 though

0.76

 admittedly

0.69

eno

0.68

 admit

0.67

tho

0.66

0.65

bie

0.64

iott

0.64

erous

0.62

 quite

0.62

Activations Density 0.061%

phrases indicating honesty or admission of truth

phrases that express honesty and admission

No Comments

No Known Activations