INDEX

Explanations

phrases where something is mentioned or identified with a specific name or term

oai_token-act-pair · gpt-3.5-turbo

instances of phrases indicating reference or attribution

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 incentive

-0.58

osion

-0.57

acity

-0.56

 Clicker

-0.56

iscopal

-0.54

redo

-0.52

 Liver

-0.52

allion

-0.52

 Frie

-0.51

 foreseeable

-0.50

POSITIVE LOGITS

 hereafter

1.04

as

0.99

 herein

0.98

 loosely

0.97

 derog

0.94

 collectively

0.93

 simply

0.89

isively

0.87

literally

0.84

 internally

0.83

Activations Density 0.078%

phrases where something is mentioned or identified with a specific name or term

instances of phrases indicating reference or attribution

No Comments

No Known Activations