INDEX

Explanations

the word "all" followed by an adjective or descriptor, indicating a comprehensive or inclusive attribute

oai_token-act-pair · gpt-3.5-turbo

phrases that include the word "all" followed by various additional descriptors

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

bledon

-0.62

IDS

-0.62

aminer

-0.61

sofar

-0.61

KH

-0.61

 Provision

-0.60

SHIP

-0.59

 Kamp

-0.59

abwe

-0.59

plin

-0.59

POSITIVE LOGITS

igator

1.25

ocating

1.17

usion

1.08

igators

1.07

uring

1.04

edged

0.99

usive

0.98

iter

0.98

ergic

0.97

iances

0.96

Activations Density 0.032%

the word "all" followed by an adjective or descriptor, indicating a comprehensive or inclusive attribute

phrases that include the word "all" followed by various additional descriptors

No Comments

No Known Activations