INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

iasis

-0.74

Muslim

-0.69

 Sabb

-0.67

Muslims

-0.65

estyles

-0.65

utions

-0.63

juven

-0.62

#$#$

-0.62

thood

-0.61

profits

-0.61

POSITIVE LOGITS

raft

0.72

ogl

0.72

orb

0.68

 profile

0.68

urst

0.68

etheus

0.67

rill

0.65

gui

0.64

assador

0.62

ryn

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.