INDEX

Explanations

the word "highest" with varying activations

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "highest" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

agra

-0.70

icer

-0.63

neg

-0.63

redo

-0.63

ca

-0.62

kas

-0.60

neys

-0.60

vous

-0.60

bal

-0.60

md

-0.59

POSITIVE LOGITS

 bidder

1.01

hest

0.85

 pinnacle

0.83

 concentrations

0.82

 rated

0.82

 extent

0.82

 imaginable

0.81

 proport

0.77

 possible

0.77

 practicable

0.77

Activations Density 0.014%

the word "highest" with varying activations

references to the concept of "highest" in various contexts

No Comments

No Known Activations