INDEX

Explanations

phrases related to specific locations or events

oai_token-act-pair · gpt-3.5-turbo

specific references to events associated with law enforcement or public safety

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

":""},{"

-0.56

theless

-0.50

pering

-0.49

oba

-0.47

=/

-0.47

omorphic

-0.46

tailed

-0.46

å¤©

-0.45

uristic

-0.45

nikov

-0.45

POSITIVE LOGITS

 universe

0.64

 saga

0.58

 empire

0.57

yssey

0.55

iverse

0.55

cosystem

0.54

 Reconstruction

0.54

senal

0.53

 presidency

0.51

 arsenal

0.51

Activations Density 0.920%

phrases related to specific locations or events

specific references to events associated with law enforcement or public safety

No Comments

No Known Activations