INDEX

Explanations

No Explanations Found

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

 relay

-0.75

OTE

-0.66

 condensed

-0.64

 eyeb

-0.62

[|

-0.62

LP

-0.61

 conclud

-0.60

IX

-0.60

ettel

-0.60

 fingerprint

-0.59

POSITIVE LOGITS

bidden

0.75

mberg

0.72

afety

0.68

steen

0.63

agascar

0.63

 Pagan

0.63

osphere

0.63

ira

0.62

rend

0.62

Atk

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.