INDEX

Explanations

mentions of the word "under" followed by a single-digit number

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "under" 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 @ 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

 Retrieved

-0.68

 Edison

-0.66

 Begins

-0.60

iterranean

-0.59

ãĥ£

-0.59

auga

-0.58

 hereafter

-0.58

 fingerprint

-0.57

andowski

-0.56

 supra

-0.56

POSITIVE LOGITS

neath

1.16

whelming

1.15

dogs

1.14

graduate

1.11

valued

1.08

stated

1.07

pants

1.05

growth

1.05

represented

1.04

lying

1.03

Activations Density 0.045%

mentions of the word "under" followed by a single-digit number

occurrences of the word "under" in various contexts

No Comments

No Known Activations