INDEX

Explanations

mentions of actions related to viewing or observing

oai_token-act-pair · gpt-3.5-turbo

instances of the word "view" and its variations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

vernment

-0.94

 whistle

-0.73

cial

-0.73

cffff

-0.73

vous

-0.71

backer

-0.71

 tackling

-0.70

wcs

-0.70

accompan

-0.69

 tightening

-0.67

POSITIVE LOGITS

ById

1.09

ports

1.05

finder

1.00

port

0.91

Į

0.81

Īè

0.78

izable

0.76

PORT

0.75

 "$:/

0.74

points

0.74

Activations Density 0.023%

mentions of actions related to viewing or observing

instances of the word "view" and its variations

No Comments

No Known Activations