INDEX

Explanations

information or situations related to connections or associations between different entities or concepts

oai_token-act-pair · gpt-3.5-turbo

references to connections or associations between concepts or entities

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

 Liberties

-0.69

Ħ¢

-0.68

ãĤ£

-0.65

»Ĵ

-0.64

sburg

-0.64

sburgh

-0.63

 actionGroup

-0.63

ribution

-0.63

XY

-0.62

perty

-0.62

POSITIVE LOGITS

edin

0.89

 thereto

0.81

 knot

0.80

 dots

0.80

 connecting

0.74

 disparate

0.73

later

0.73

linked

0.69

 dise

0.69

 link

0.68

Activations Density 0.033%

information or situations related to connections or associations between different entities or concepts

references to connections or associations between concepts or entities

No Comments

No Known Activations