INDEX

Explanations

names of individuals related to news or politics

oai_token-act-pair · gpt-3.5-turbo

mentions of specific individuals, particularly "Maddow" and "Maggie"

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 microwave

-0.74

ccording

-0.72

 Origin

-0.71

ASED

-0.69

pta

-0.64

 Viking

-0.64

 predatory

-0.64

âĸ¬

-0.64

chnology

-0.64

 SCHOOL

-0.63

POSITIVE LOGITS

 Madd

1.11

ox

0.92

sylv

0.91

oline

0.90

enh

0.90

ota

0.89

ison

0.85

ani

0.85

alos

0.85

ings

0.83

Activations Density 0.007%

names of individuals related to news or politics

mentions of specific individuals, particularly "Maddow" and "Maggie"

No Comments

No Known Activations