INDEX

Explanations

dates and events

oai_token-act-pair · gpt-3.5-turbo

occurrences of parentheses

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

 flock

-0.76

wre

-0.71

 escaping

-0.70

 intensive

-0.70

 appropri

-0.69

 defect

-0.66

 Emin

-0.66

 evapor

-0.66

 organic

-0.66

 leveling

-0.66

POSITIVE LOGITS

pictured

1.45

sic

1.34

CBC

1.19

emphasis

1.13

AFP

1.12

see

1.11

PDF

1.10

pdf

1.08

formerly

1.08

Reuters

1.08

Activations Density 0.176%

dates and events

occurrences of parentheses

No Comments

No Known Activations