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

selves

-0.80

itatively

-0.76

isite

-0.75

alties

-0.72

ideos

-0.69

ibrary

-0.68

RIPT

-0.68

Assistant

-0.67

Paper

-0.67

roads

-0.67

POSITIVE LOGITS

Rating

0.71

 clamp

0.68

 burger

0.64

 squid

0.64

 lunch

0.64

 crunch

0.62

pir

0.61

 pant

0.61

 Thai

0.61

 fridge

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.