INDEX

Explanations

references to winning or victory in electoral contexts

oai_token-act-pair · gpt-3.5-turbo

instances of the word "won" and its variations, particularly in the context of elections

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

uras

-0.76

IDER

-0.72

ĸļ

-0.68

 Minotaur

-0.66

 pros

-0.64

 ling

-0.64

eco

-0.64

Factor

-0.62

 relie

-0.62

 deprecated

-0.62

POSITIVE LOGITS

 reelection

1.08

 majorities

0.96

 votes

0.92

elect

0.90

 elections

0.89

election

0.89

 elector

0.89

 landslide

0.89

't

0.88

 electoral

0.85

Activations Density 0.083%

references to winning or victory in electoral contexts

instances of the word "won" and its variations, particularly in the context of elections

No Comments

No Known Activations