INDEX

Explanations

phrases related to descriptions or accounts given by eyewitnesses

oai_token-act-pair · gpt-3.5-turbo

references to eyewitness accounts and testimonies

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

ono

-0.72

atism

-0.71

rost

-0.69

aptop

-0.69

ategory

-0.69

oku

-0.69

pta

-0.68

ojure

-0.67

akings

-0.67

appa

-0.66

POSITIVE LOGITS

ewitness

1.47

Witness

1.11

 eyewitness

0.97

 testimonies

0.94

 witnesses

0.90

itness

0.89

 testimony

0.87

 witness

0.80

 Witnesses

0.79

 bystand

0.71

Activations Density 0.013%

phrases related to descriptions or accounts given by eyewitnesses

references to eyewitness accounts and testimonies

No Comments

No Known Activations