INDEX

Explanations

locations or landmarks, especially mountains and volcanoes

oai_token-act-pair · gpt-3.5-turbo

names of geographical locations, particularly mountains and forts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

organic

-0.68

inant

-0.65

staking

-0.63

iegel

-0.63

itled

-0.62

 Spiegel

-0.61

 Hyundai

-0.61

vertising

-0.61

 arrang

-0.60

abase

-0.59

POSITIVE LOGITS

 Reserv

0.81

burgh

0.78

llan

0.74

arde

0.74

 Island

0.73

 Despair

0.73

 Islands

0.71

helm

0.71

 Cemetery

0.70

obyl

0.70

Activations Density 0.108%

locations or landmarks, especially mountains and volcanoes

names of geographical locations, particularly mountains and forts

No Comments

No Known Activations