INDEX

Explanations

specific references to locations or time frames

oai_token-act-pair · gpt-3.5-turbo

instances of the word "that."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Attacks

-0.79

å§«

-0.74

izons

-0.74

ocks

-0.73

amps

-0.73

cycles

-0.72

bugs

-0.71

irds

-0.71

zos

-0.70

okers

-0.70

POSITIVE LOGITS

 fateful

1.50

 same

1.33

 particular

1.26

 pesky

1.08

 elusive

1.02

 infamous

1.01

 exact

0.98

ched

0.97

 dreadful

0.95

 awful

0.93

Activations Density 0.101%

specific references to locations or time frames

instances of the word "that."

No Comments

No Known Activations