INDEX

Explanations

references to specific items or objects, such as names, locations, and identifiers

oai_token-act-pair · gpt-3.5-turbo

references to specific entities or objects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Canaver

-0.79

forts

-0.78

venants

-0.74

doms

-0.74

cosystem

-0.73

achus

-0.73

folk

-0.70

 havens

-0.70

izons

-0.70

agy

-0.69

POSITIVE LOGITS

 offending

0.91

 original

0.90

 previous

0.88

 latter

0.87

 phone

0.86

 actual

0.85

 item

0.84

 desired

0.84

 deed

0.83

 receipt

0.82

Activations Density 0.513%

references to specific items or objects, such as names, locations, and identifiers

references to specific entities or objects

No Comments

No Known Activations