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

ience

-0.71

rooms

-0.71

ele

-0.70

URN

-0.70

orough

-0.69

locks

-0.69

aken

-0.68

ric

-0.68

raid

-0.67

rol

-0.66

POSITIVE LOGITS

 West

0.81

 Terminator

0.70

 Manhattan

0.63

 Heights

0.63

 Bash

0.63

 Andromeda

0.62

Ug

0.62

velt

0.62

hof

0.61

 Democr

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.