INDEX

Explanations

expressions related to threats and conflict

oai_token-act-pair · gpt-3.5-turbo

expressions of conflict or threat

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 lately

-0.71

amiliar

-0.59

lyak

-0.59

erd

-0.56

artney

-0.54

bered

-0.54

emis

-0.54

Enlarge

-0.54

itled

-0.54

 Entered

-0.53

POSITIVE LOGITS

if

1.64

 anytime

1.39

 unless

1.38

 whenever

1.32

 someday

1.30

 whoever

1.26

 whichever

1.26

 wherever

1.24

 hereafter

1.16

unless

1.12

Activations Density 0.729%

expressions related to threats and conflict

expressions of conflict or threat

No Comments

No Known Activations