INDEX

Explanations

proper nouns, specifically names

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of the name "Es."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 ration

-0.84

 Nadu

-0.73

 rounding

-0.69

ahead

-0.63

itiveness

-0.63

atically

-0.63

 bothered

-0.62

anmar

-0.62

 bamboo

-0.62

 crackdown

-0.61

POSITIVE LOGITS

Es

3.96

Es

2.19

es

1.68

Eb

1.19

Os

1.18

ES

1.18

Osc

1.18

Eps

1.18

Esc

1.15

Ez

1.10

Activations Density 0.008%

proper nouns, specifically names

repeated mentions of the name "Es."

No Comments

No Known Activations