INDEX

Explanations

mentions of geographical locations, specifically countries and regions

oai_token-act-pair · gpt-3.5-turbo

references to the country of Panama and related geographical terms

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

chen

-0.87

ä

-0.81

itbart

-0.80

ando

-0.79

eting

-0.76

share

-0.74

places

-0.74

itar

-0.73

chel

-0.73

¨

-0.71

POSITIVE LOGITS

 Rican

0.96

 reefs

0.91

Sea

0.85

 Plain

0.84

 Islands

0.82

 Carib

0.80

 Rica

0.76

 Beach

0.75

 Strait

0.74

 Island

0.74

Activations Density 0.055%

mentions of geographical locations, specifically countries and regions

references to the country of Panama and related geographical terms

No Comments

No Known Activations