INDEX

Explanations

mentions of specific locations or entities, especially universities and medical terms with a possible country context

oai_token-act-pair · gpt-3.5-turbo

references to specific locations or regions

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

eson

-0.80

atile

-0.78

hester

-0.77

ively

-0.77

otto

-0.74

esc

-0.74

es

-0.74

atively

-0.73

enegger

-0.68

umin

-0.68

POSITIVE LOGITS

 Skydragon

0.86

ocracy

0.83

ulhu

0.77

knife

0.74

breaker

0.74

Bay

0.73

pta

0.73

assian

0.72

 Concord

0.72

ffen

0.72

Activations Density 0.028%

mentions of specific locations or entities, especially universities and medical terms with a possible country context

references to specific locations or regions

No Comments

No Known Activations