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

hardt

-0.79

ellen

-0.79

HAEL

-0.73

senal

-0.72

gur

-0.71

ush

-0.68

ãĤ·ãĥ£

-0.68

wana

-0.68

 Alive

-0.68

holm

-0.67

POSITIVE LOGITS

DEBUG

0.68

ovo

0.66

 proposition

0.65

ithering

0.63

operated

0.61

ults

0.60

atform

0.60

oran

0.59

 footprints

0.59

 dictatorship

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.