INDEX

Explanations

proper nouns and names, potentially related to locations or entities

oai_token-act-pair · gpt-3.5-turbo

punctuation, specifically commas and periods

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

zing

-0.68

¦

-0.66

worldly

-0.66

¯

-0.66

:[

-0.65

ãĥ¼

-0.63

ĸ

-0.63

verage

-0.62

Ĳ

-0.60

't

-0.60

POSITIVE LOGITS

 dominates

1.04

 arrives

1.00

 appears

0.99

 became

0.97

was

0.97

has

0.97

 belongs

0.97

 appeared

0.96

 resides

0.96

 earns

0.96

Activations Density 0.292%

proper nouns and names, potentially related to locations or entities

punctuation, specifically commas and periods

No Comments

No Known Activations