INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 Hansen

-0.74

 Somebody

-0.66

 Harden

-0.66

rou

-0.65

ilater

-0.62

 Krug

-0.62

lessly

-0.60

 overrun

-0.60

ilar

-0.59

 Goldberg

-0.58

POSITIVE LOGITS

wcsstore

0.66

 behav

0.65

baugh

0.64

BuyableInstoreAndOnline

0.64

Beg

0.64

CLASSIFIED

0.63

edin

0.63

Cola

0.62

ktop

0.61

Interested

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.