INDEX

Explanations

phrases related to a particular political party named "Lib" with varying activation values

oai_token-act-pair · gpt-3.5-turbo

references to the Liberal Democrats (Lib Dems)

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

theless

-0.69

 Order

-0.69

 silence

-0.68

 boarding

-0.67

 Oaks

-0.67

ufact

-0.65

OPLE

-0.65

Ħ¢

-0.64

 Athletic

-0.63

Zed

-0.61

POSITIVE LOGITS

virt

1.13

ido

0.99

tex

0.94

rium

0.90

bers

0.89

aries

0.87

rum

0.86

bey

0.85

ris

0.85

syn

0.85

Activations Density 0.032%

phrases related to a particular political party named "Lib" with varying activation values

references to the Liberal Democrats (Lib Dems)

No Comments

No Known Activations