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

 "$:/

-0.77

rils

-0.68

 tongues

-0.68

 announ

-0.67

bent

-0.66

 tipped

-0.66

 purs

-0.66

 archaeologists

-0.65

 organisers

-0.64

 tong

-0.64

POSITIVE LOGITS

ullivan

0.84

ãĥ¼ãĥĨ

0.77

 msec

0.71

TB

0.70

DX

0.69

iant

0.69

gb

0.69

qus

0.68

ths

0.68

 unfocusedRange

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.