INDEX

Explanations

What best represents what the neuron is looking for is names or terms containing 'gl'

oai_token-act-pair · gpt-3.5-turbo

suffixes related to locations or geographical names

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

hift

-0.70

flix

-0.68

ldon

-0.68

netflix

-0.67

ãĤ´ãĥ³

-0.66

laws

-0.63

creen

-0.62

Reviewer

-0.62

goers

-0.62

 Rowling

-0.61

POSITIVE LOGITS

iano

1.10

iani

1.00

obal

0.91

otte

0.87

estial

0.85

ott

0.83

iffe

0.83

ius

0.83

agne

0.82

estine

0.77

Activations Density 0.077%

What best represents what the neuron is looking for is names or terms containing 'gl'

suffixes related to locations or geographical names

No Comments

No Known Activations