INDEX

Explanations

phrases or terms related to positioning or placement

oai_token-act-pair · gpt-3.5-turbo

instances of the word "lay"

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

ilar

-0.77

andom

-0.69

GB

-0.66

illas

-0.66

entric

-0.65

ilee

-0.65

okemon

-0.65

illa

-0.62

qua

-0.62

POSITIVE LOGITS

 showc

0.88

sheets

0.88

 arrang

0.87

\\\\\\\\

0.86

sembly

0.86

 newcom

0.84

lay

0.84

uten

0.83

ayer

0.82

 theoret

0.81

Activations Density 0.005%

phrases or terms related to positioning or placement

instances of the word "lay"

No Comments

No Known Activations