INDEX

Explanations

statements related to research findings or studies

oai_token-act-pair · gpt-3.5-turbo

discussions about scientific studies and their conclusions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

orld

-0.74

tesy

-0.74

ruary

-0.74

@#

-0.73

yssey

-0.73

çīĪ

-0.73

=-=-

-0.73

\\\\\\\\

-0.73

Tokens

-0.72

nesday

-0.71

POSITIVE LOGITS

 obesity

1.23

 antidepressant

1.14

 antibiotic

1.12

 contraceptive

1.10

 antidepressants

1.09

HPV

1.06

 genetically

1.05

 diets

1.05

 fertility

1.05

 cancers

1.04

Activations Density 0.756%

statements related to research findings or studies

discussions about scientific studies and their conclusions

No Comments

No Known Activations