INDEX

Explanations

phrases related to nonpartisanship and impartiality

oai_token-act-pair · gpt-3.5-turbo

terms related to nonpartisan entities and impartiality in discussions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

Interstitial

-0.82

Pac

-0.74

KT

-0.71

 Typhoon

-0.71

 Typh

-0.70

Phones

-0.69

Yar

-0.68

 Bulg

-0.68

 Hughes

-0.67

Ku

-0.67

POSITIVE LOGITS

 nonpartisan

1.12

 impartial

1.00

partisan

0.93

vernment

0.89

artisan

0.88

sonian

0.87

 unbiased

0.78

 observer

0.77

 watchdog

0.75

ervative

0.75

Activations Density 0.015%

phrases related to nonpartisanship and impartiality

terms related to nonpartisan entities and impartiality in discussions

No Comments

No Known Activations