INDEX

Explanations

terms related to accessibility and technological features

oai_token-act-pair · gpt-3.5-turbo

references to access and accessibility issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Kern

-0.80

astically

-0.66

annis

-0.64

Rag

-0.61

 ejac

-0.61

Lens

-0.60

 Saud

-0.59

 bulls

-0.59

 sperm

-0.58

 abst

-0.58

POSITIVE LOGITS

ibility

1.50

ibly

1.24

ible

1.19

ory

1.09

ibles

1.09

ories

1.08

ibl

1.03

ibilities

0.96

IBLE

0.95

ioned

0.91

Activations Density 0.047%

terms related to accessibility and technological features

references to access and accessibility issues

No Comments

No Known Activations