INDEX

Explanations

mentions of celebrities

oai_token-act-pair · gpt-3.5-turbo

mentions of celebrities

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

choes

-0.84

anus

-0.81

hematic

-0.80

tered

-0.76

nea

-0.73

doors

-0.73

nerg

-0.72

sterdam

-0.71

THER

-0.70

¼

-0.68

POSITIVE LOGITS

rities

1.16

 endors

1.01

 endorsements

1.00

 celebrities

0.90

 chef

0.88

wcs

0.87

 chefs

0.85

 gossip

0.83

 idols

0.80

 celebrity

0.79

Activations Density 0.021%

mentions of celebrities

mentions of celebrities

No Comments

No Known Activations