INDEX

Explanations

words related to technology, concepts, and industry

oai_token-act-pair · gpt-3.5-turbo

punctuation or commas used to delineate thoughts or clauses

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

tsky

-0.86

enf

-0.76

lash

-0.73

panic

-0.72

resa

-0.71

lift

-0.71

izens

-0.71

enary

-0.70

 Moines

-0.70

inar

-0.70

POSITIVE LOGITS

 albeit

1.36

 although

1.12

 akin

1.11

 consisting

1.08

 comprising

1.06

 combining

1.05

 wherein

1.04

 embod

1.00

 which

1.00

 requiring

1.00

Activations Density 0.331%

words related to technology, concepts, and industry

punctuation or commas used to delineate thoughts or clauses

No Comments

No Known Activations