INDEX

Explanations

terms related to books or book titles

oai_token-act-pair · gpt-3.5-turbo

references to specific books

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

ilitary

-0.86

asio

-0.84

Downloadha

-0.80

 corrosion

-0.75

 VIDEOS

-0.71

asonic

-0.70

 sclerosis

-0.69

 territ

-0.68

orescence

-0.67

GoldMagikarp

-0.67

POSITIVE LOGITS

stores

1.32

book

1.19

marks

1.13

mark

1.12

store

1.10

worm

1.09

shop

1.05

seller

1.04

cases

1.02

Book

1.00

Activations Density 0.030%

terms related to books or book titles

references to specific books

No Comments

No Known Activations