INDEX

Explanations

locations or points of interest related to specific towns or entities

oai_token-act-pair · gpt-3.5-turbo

occurrences of proper names and specific identifiers

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

ħĭ

-0.81

querque

-0.79

ãĥ¼ãĥĨ

-0.75

vantage

-0.75

inally

-0.71

 DEFENSE

-0.70

£ı

-0.69

INAL

-0.68

regor

-0.65

rities

-0.63

POSITIVE LOGITS

nesday

1.02

ipeg

0.97

ÅĤ

0.95

atche

0.81

itzer

0.81

bledon

0.80

hester

0.79

lich

0.78

wright

0.77

atever

0.74

Activations Density 0.107%

locations or points of interest related to specific towns or entities

occurrences of proper names and specific identifiers

No Comments

No Known Activations