INDEX

Explanations

hyperlinks or calls to action in a text

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "HERE" in various contexts

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

gling

-0.80

bath

-0.69

egu

-0.68

utter

-0.67

omore

-0.67

ible

-0.65

urger

-0.64

usher

-0.64

eff

-0.63

arma

-0.63

POSITIVE LOGITS

 HERE

1.12

NOW

0.90

 THERE

0.89

 BELOW

0.88

FORE

0.85

WHERE

0.85

 INCLUD

0.81

 NEXT

0.81

 PHOTO

0.81

tical

0.80

Activations Density 0.010%

hyperlinks or calls to action in a text

occurrences of the word "HERE" in various contexts

No Comments

No Known Activations