INDEX

Explanations

phrases indicating written communication

oai_token-act-pair · gpt-3.5-turbo

instances of the word "wrote" and various related phrases

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 ILCS

-0.76

agara

-0.75

phant

-0.72

ĪĴ

-0.70

Desktop

-0.70

Ĭ±

-0.68

Unity

-0.68

AMI

-0.67

illon

-0.66

anu

-0.65

POSITIVE LOGITS

 extensively

0.90

 aloud

0.87

 eloqu

0.87

 letters

0.87

 scathing

0.82

 poems

0.80

 about

0.79

 sarcast

0.78

 Letters

0.78

 blog

0.77

Activations Density 0.039%

phrases indicating written communication

instances of the word "wrote" and various related phrases

No Comments

No Known Activations