INDEX

Explanations

mentions of specific words such as "Bloom", "Frost", and "Steele"

oai_token-act-pair · gpt-3.5-turbo

specific names and references to people and brands

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

alg

-0.81

rupulous

-0.81

paralle

-0.78

alien

-0.75

haps

-0.74

igation

-0.73

terior

-0.73

arnaev

-0.72

asks

-0.71

gettable

-0.70

POSITIVE LOGITS

field

0.88

 Bloom

0.84

heim

0.79

fields

0.78

berries

0.76

burn

0.75

berry

0.74

 Frost

0.74

Burn

0.74

 Payne

0.73

Activations Density 0.010%

mentions of specific words such as "Bloom", "Frost", and "Steele"

specific names and references to people and brands

No Comments

No Known Activations