INDEX

Explanations

complex phrases or clauses beginning with "which"

oai_token-act-pair · gpt-3.5-turbo

instances of the word "which", often indicating elaborate or descriptive clauses

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

Bed

-0.74

 Baker

-0.70

politics

-0.66

 Bucc

-0.66

let

-0.65

Behind

-0.62

 Tracker

-0.62

Caf

-0.62

Ott

-0.60

Et

-0.60

POSITIVE LOGITS

soever

1.00

xual

0.86

 resulted

0.82

 exposes

0.79

allows

0.79

 brings

0.77

ĸļ

0.76

 attracts

0.76

 enables

0.75

amar

0.74

Activations Density 0.114%

complex phrases or clauses beginning with "which"

instances of the word "which", often indicating elaborate or descriptive clauses

No Comments

No Known Activations