INDEX

Explanations

references to people speaking or giving speeches

oai_token-act-pair · gpt-3.5-turbo

instances of the word "speaking"

oai_token-act-pair · gpt-4o-mini Triggered by @bot

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

strip

-0.72

ascade

-0.71

-0.70

\/\/

-0.67

gradation

-0.67

utsu

-0.66

urses

-0.66

ologne

-0.66

aredevil

-0.65

jer

-0.64

POSITIVE LOGITS

 aloud

0.96

 engagements

0.90

 fluent

0.86

 voices

0.83

ophone

0.80

 louder

0.77

 voice

0.74

 speakers

0.74

voice

0.73

 volumes

0.73

Activations Density 0.012%

references to people speaking or giving speeches

instances of the word "speaking"

No Comments

No Known Activations