INDEX

Explanations

keywords related to providing additional information, details, options, or assistance

oai_token-act-pair · gpt-3.5-turbo

references to extra resources or supplementary content

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

dog

-0.83

bane

-0.80

bows

-0.79

 Alone

-0.69

ESE

-0.69

 Diet

-0.68

 Nose

-0.67

 Â¯

-0.65

generic

-0.64

 Laden

-0.64

POSITIVE LOGITS

 additional

0.91

requisite

0.89

judicial

0.84

ãĤī

0.83

 bonuses

0.81

 additions

0.80

 precaution

0.79

 disclosure

0.79

elaide

0.79

 surprises

0.79

Activations Density 0.012%

keywords related to providing additional information, details, options, or assistance

references to extra resources or supplementary content

No Comments

No Known Activations