INDEX

Explanations

words related to formal decision-making processes, particularly involving voting

oai_token-act-pair · gpt-3.5-turbo

instances of the word "voted" and related voting actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

fal

-0.86

ãĤ¨ãĥ«

-0.79

ahime

-0.72

 unden

-0.72

mone

-0.68

thur

-0.68

abol

-0.67

 millenn

-0.66

illin

-0.66

sie

-0.66

POSITIVE LOGITS

 votes

1.08

 voted

0.97

 Votes

0.96

 vote

0.95

 Vote

0.94

 unanimously

0.93

 voting

0.92

vote

0.89

 ballots

0.85

ndum

0.81

Activations Density 0.024%

words related to formal decision-making processes, particularly involving voting

instances of the word "voted" and related voting actions

No Comments

No Known Activations