INDEX

Explanations

statements or declarations in text

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

elsius

-0.83

cffff

-0.79

NetMessage

-0.77

rys

-0.73

iencies

-0.72

sites

-0.71

itals

-0.70

rowd

-0.69

nown

-0.68

 Pradesh

-0.68

POSITIVE LOGITS

 regarding

0.96

 condemning

0.95

 announcing

0.92

 statement

0.89

 denouncing

0.89

 statements

0.86

 concerning

0.85

 apologizing

0.83

 proclaiming

0.83

 issued

0.83

Activations Density 0.059%

statements or declarations in text

No Comments

No Known Activations