INDEX

Explanations

phrases related to people's work history

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "worked" 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 @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 constitu

-0.70

Andre

-0.65

ĪĴ

-0.65

emonic

-0.63

idium

-0.62

ylon

-0.62

 Augustus

-0.62

xual

-0.61

IX

-0.61

iren

-0.60

POSITIVE LOGITS

 tirelessly

1.03

hops

0.97

 diligently

0.97

 ethic

0.96

hirt

0.93

paces

0.90

 collabor

0.89

forces

0.83

heet

0.82

aday

0.80

Activations Density 0.067%

phrases related to people's work history

occurrences of the word "worked" in various contexts

No Comments

No Known Activations