INDEX

Explanations

statements of advocacy or recommendation

oai_token-act-pair · gpt-3.5-turbo

calls to action or requests for action

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

çĦ

-0.75

ëĭ

-0.67

ìĿ

-0.64

central

-0.64

sf

-0.64

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.61

á

-0.61

Û

-0.61

buck

-0.60

 Antiqu

-0.59

POSITIVE LOGITS

 caution

0.92

 pressuring

0.79

wcs

0.78

irection

0.75

 urging

0.74

urous

0.73

 prompt

0.73

FontSize

0.72

tale

0.71

 anew

0.71

Activations Density 0.021%

statements of advocacy or recommendation

calls to action or requests for action

No Comments

No Known Activations