INDEX

Explanations

references to positions of authority or leadership roles

oai_token-act-pair · gpt-3.5-turbo

references to individuals with the title "chief" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

phis

-0.88

nikov

-0.83

ajor

-0.80

ppo

-0.79

aughs

-0.78

lished

-0.75

etimes

-0.74

avery

-0.70

ichen

-0.70

areth

-0.69

POSITIVE LOGITS

 executive

1.04

doms

1.00

iary

0.86

 executives

0.83

 Executive

0.83

 negotiator

0.82

stay

0.81

IAL

0.80

tenance

0.74

 contributor

0.74

Activations Density 0.025%

references to positions of authority or leadership roles

references to individuals with the title "chief" in various contexts

No Comments

No Known Activations