INDEX

Explanations

company names and product references

oai_token-act-pair · gpt-3.5-turbo

references to organizations and entities in a political context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

nell

-1.09

ility

-0.95

glers

-0.83

bos

-0.82

ilities

-0.78

auts

-0.76

icles

-0.76

sonian

-0.76

swick

-0.75

ãĤ£

-0.73

POSITIVE LOGITS

TextColor

0.86

endra

0.77

 earthquake

0.73

sr

0.68

 mosqu

0.68

ikuman

0.67

Subject

0.65

ä¿

0.63

 Penalty

0.63

 Lumpur

0.62

Activations Density 0.024%

company names and product references

references to organizations and entities in a political context

No Comments

No Known Activations