INDEX

Explanations

mentions of or related to humans

oai_token-act-pair · gpt-3.5-turbo

occurrences and discussions of humans

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Line

-0.67

rent

-0.67

forth

-0.67

Coun

-0.67

abb

-0.65

PM

-0.64

acc

-0.63

 Stud

-0.62

 Sutherland

-0.62

 Scarborough

-0.61

POSITIVE LOGITS

 beings

1.18

 humans

0.97

 Humans

0.96

folk

0.96

 readable

0.82

omorphic

0.82

oids

0.81

 mortals

0.76

izont

0.75

zee

0.72

Activations Density 0.012%

mentions of or related to humans

occurrences and discussions of humans

No Comments

No Known Activations