INDEX

Explanations

technical terms and jargon related to information technology and scientific research

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

Data Type

torch.float32

Hook Point

blocks.8.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

 tender

-0.75

bid

-0.72

 duel

-0.72

 targets

-0.72

 avail

-0.71

 racket

-0.70

 fathers

-0.70

 poaching

-0.70

 speeches

-0.69

 glide

-0.69

POSITIVE LOGITS

etc

1.56

whatever

1.32

Flickr

1.31

ãĤ¨ãĥ«

1.19

Pg

1.14

CN

1.12

CIA

1.11

acqu

1.11

1.09

oct

1.08

Activations Density 0.230%

technical terms and jargon related to information technology and scientific research

No Comments

No Known Activations