INDEX

Explanations

specific mentions of speeches

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Users

-0.66

âĺĨ

-0.63

ccording

-0.61

 intrusive

-0.59

 Soldiers

-0.58

finder

-0.58

 Associ

-0.58

VIDIA

-0.57

 Restaur

-0.57

 crossings

-0.57

POSITIVE LOGITS

writers

1.26

writer

1.22

writing

1.09

 delivered

0.97

 speech

0.90

 speeches

0.90

 transcripts

0.88

 denouncing

0.87

speech

0.87

clip

0.86

Activations Density 0.094%

specific mentions of speeches

No Comments

No Known Activations