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

 tang

-0.82

 References

-0.75

 principled

-0.65

 therapy

-0.65

 restitution

-0.64

 proclamation

-0.64

 Kardash

-0.61

 implication

-0.61

Comments

-0.60

 consequential

-0.60

POSITIVE LOGITS

heon

0.75

essor

0.72

yip

0.72

ahon

0.69

gun

0.68

umn

0.67

auga

0.67

ombat

0.66

umbnail

0.65

ioxide

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.