INDEX

Explanations

proper nouns, specifically names of political figures and locations

oai_token-act-pair · gpt-3.5-turbo

variations of the word "assess"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

istically

-0.78

 glam

-0.65

ests

-0.65

isher

-0.62

iggins

-0.62

ised

-0.61

olls

-0.61

opsy

-0.61

spring

-0.61

 opener

-0.60

POSITIVE LOGITS

asse

1.12

ment

0.79

uve

0.79

ments

0.78

bec

0.76

âĹ¼

0.76

 Hollande

0.75

xual

0.73

rand

0.72

uble

0.72

Activations Density 0.012%

proper nouns, specifically names of political figures and locations

variations of the word "assess"

No Comments

No Known Activations