INDEX

Explanations

words related to achievements or accomplishments

oai_token-act-pair · gpt-3.5-turbo

mentions of universities and academic contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 synopsis

-0.71

Gem

-0.65

deal

-0.63

 saliva

-0.62

 technicians

-0.62

 opinion

-0.61

 Trailer

-0.61

gem

-0.61

 spoilers

-0.60

 Clean

-0.60

POSITIVE LOGITS

arsity

3.71

iking

2.88

iked

1.40

oodoo

1.31

intage

1.21

aunted

1.13

ebra

1.12

anguard

1.09

itals

1.01

ocation

0.98

Activations Density 0.031%

words related to achievements or accomplishments

mentions of universities and academic contexts

No Comments

No Known Activations