INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ãĤ¦ãĤ¹

-0.76

itiveness

-0.68

ãĥ¼ãĥ«

-0.65

ãĥĥãĥĪ

-0.63

Magikarp

-0.62

 gratification

-0.62

Rating

-0.60

 Rasm

-0.60

 Straw

-0.60

maxwell

-0.60

POSITIVE LOGITS

lege

0.77

authorized

0.72

reditary

0.71

rosso

0.67

trained

0.67

arus

0.66

auga

0.64

inia

0.64

nia

0.64

 banned

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.