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

stellar

-0.81

wire

-0.74

ilver

-0.73

vo

-0.72

aryl

-0.71

evin

-0.69

ippi

-0.67

need

-0.67

draft

-0.65

uliffe

-0.64

POSITIVE LOGITS

 Slayer

0.72

âĦ¢:

0.69

GI

0.69

Sov

0.64

Magikarp

0.64

 Vaughan

0.63

icz

0.63

 Buddh

0.61

psy

0.60

 steroids

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.