INDEX

Explanations

numbers indicating a specific ranking or score in a range from 0 to 10

oai_token-act-pair · gpt-3.5-turbo

numeric values and references to measurements or statistics

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

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

netflix

-0.84

EStream

-0.76

rette

-0.72

TextColor

-0.72

deen

-0.71

ramid

-0.67

raviolet

-0.67

sole

-0.66

 propriet

-0.66

quet

-0.65

POSITIVE LOGITS

0.93

0.88

^{

0.88

0.86

0.85

acity

0.84

Activations Density 0.060%

numbers indicating a specific ranking or score in a range from 0 to 10

numeric values and references to measurements or statistics

No Comments

No Known Activations