INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

hester

-0.80

 Surviv

-0.78

imum

-0.77

leon

-0.72

 Reverse

-0.69

pmwiki

-0.67

 Blaze

-0.67

 specificity

-0.66

ãĤ¤ãĥĪ

-0.66

å§«

-0.64

POSITIVE LOGITS

 guiActiveUn

0.66

opian

0.63

elling

0.61

idges

0.61

 pessimistic

0.60

rief

0.60

obil

0.59

itamin

0.59

 Moody

0.58

 reflective

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.