INDEX

Explanations

words related to taking actions or causing actions to happen

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate actions or processes involving "to" as a preposition

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 preceded

-0.69

 evid

-0.66

few

-0.62

 damned

-0.61

 casing

-0.61

staking

-0.61

Few

-0.60

 Cust

-0.60

 inherited

-0.60

Disclaimer

-0.60

POSITIVE LOGITS

 grips

1.00

asted

0.87

tten

0.78

encers

0.78

ggles

0.76

ullivan

0.74

ppers

0.71

plet

0.70

 cooperate

0.70

 meet

0.70

Activations Density 0.094%

words related to taking actions or causing actions to happen

phrases that indicate actions or processes involving "to" as a preposition

No Comments

No Known Activations