INDEX

Explanations

proper nouns, specifically names of individuals

oai_token-act-pair · gpt-3.5-turbo

specific names or references to individuals, particularly those related to legal or official 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

cloth

-0.74

adesh

-0.67

FN

-0.66

fit

-0.65

unit

-0.64

riors

-0.64

Site

-0.63

ported

-0.63

ichita

-0.63

Corp

-0.62

POSITIVE LOGITS

xual

0.80

ennial

0.73

icles

0.69

 Clement

0.68

inho

0.68

Nir

0.66

 pulp

0.65

RAL

0.64

obal

0.64

ipation

0.63

Activations Density 0.036%

proper nouns, specifically names of individuals

specific names or references to individuals, particularly those related to legal or official contexts

No Comments

No Known Activations