INDEX

Explanations

text related to educational resources

oai_token-act-pair · gpt-3.5-turbo

terms related to resources and information sources

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 milliseconds

-0.80

conn

-0.77

engeance

-0.70

 forgiven

-0.66

inction

-0.66

 externalToEVAOnly

-0.66

 forfeit

-0.64

 revenge

-0.64

 unlucky

-0.64

 comeback

-0.62

POSITIVE LOGITS

 guides

1.24

 informative

1.06

 Guides

1.04

 guide

1.03

 appendix

0.99

endium

0.99

guide

0.98

 Guide

0.97

 instructional

0.97

pdf

0.94

Activations Density 0.483%

text related to educational resources

terms related to resources and information sources

No Comments

No Known Activations