INDEX

Explanations

proper nouns or names, possibly related to news or events

oai_token-act-pair · gpt-3.5-turbo

the letter "G" in various contexts

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

ĸļ

-0.89

terday

-0.64

 compr

-0.63

oÄŁ

-0.62

ĺħ

-0.62

 derail

-0.61

 justified

-0.61

 retrospect

-0.60

 provision

-0.59

Bie

-0.59

POSITIVE LOGITS

roups

1.28

iants

1.21

raphic

1.17

uild

1.11

iant

1.06

ossip

1.06

asp

1.05

AMES

1.05

reetings

1.04

RAY

1.03

Activations Density 0.038%

proper nouns or names, possibly related to news or events

the letter "G" in various contexts

No Comments

No Known Activations