INDEX

Explanations

proper nouns

oai_token-act-pair · gpt-3.5-turbo

references to the word "Ba" followed by various numerical identifiers or context descriptors

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

pants

-0.79

lessly

-0.76

wise

-0.74

otle

-0.68

 address

-0.67

REDACTED

-0.67

ments

-0.64

dress

-0.60

weed

-0.60

 Leone

-0.60

POSITIVE LOGITS

uble

1.20

uman

1.14

iley

1.12

umann

1.11

plin

1.07

iting

1.03

atar

1.01

ÅŁ

0.99

Ã±

0.99

um

0.97

Activations Density 0.026%

proper nouns

references to the word "Ba" followed by various numerical identifiers or context descriptors

No Comments

No Known Activations