INDEX

Explanations

political terms and entities

oai_token-act-pair · gpt-3.5-turbo

terms related to politics and governance

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

DragonMagazine

-0.71

Ô

-0.65

awaru

-0.63

govtrack

-0.62

theless

-0.59

å§«

-0.58

sbm

-0.56

terday

-0.56

 Edited

-0.55

ertodd

-0.55

POSITIVE LOGITS

iest

0.95

liest

0.82

osphere

0.82

est

0.77

 portion

0.72

era

0.71

 side

0.70

hest

0.70

 realm

0.67

 arena

0.66

Activations Density 1.043%

political terms and entities

terms related to politics and governance

No Comments

No Known Activations