INDEX

Explanations

references to political entities, specifically republics

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "republic" and related terms

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

Ack

-0.81

gaard

-0.77

 Ammunition

-0.74

itton

-0.74

 Kamp

-0.69

STD

-0.67

 Petersen

-0.67

acha

-0.63

velength

-0.63

Oracle

-0.63

POSITIVE LOGITS

lisher

0.99

0.98

rats

0.95

hips

0.94

rants

0.84

inals

0.83

teen

0.80

sie

0.80

etrical

0.79

sis

0.79

Activations Density 0.038%

references to political entities, specifically republics

occurrences of the word "republic" and related terms

No Comments

No Known Activations