INDEX

Explanations

locations or places, especially areas or regions

oai_token-act-pair · gpt-3.5-turbo

names of geographic 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 @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

tyard

-0.81

ynt

-0.77

odox

-0.75

neapolis

-0.73

kaya

-0.71

wik

-0.71

Bir

-0.70

onymous

-0.70

VALUE

-0.70

Module

-0.70

POSITIVE LOGITS

 Edison

0.85

 Territories

0.80

 RCMP

0.78

 Militia

0.78

 winters

0.77

 natives

0.76

 Territory

0.76

 wildfires

0.75

 Presbyterian

0.75

 Correctional

0.74

Activations Density 0.158%

locations or places, especially areas or regions

names of geographic locations or regions

No Comments

No Known Activations