INDEX

Explanations

names of political figures

oai_token-act-pair · gpt-3.5-turbo

occurrences of specific names, particularly individuals and places

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

eker

-0.74

uary

-0.71

oxide

-0.70

 Skydragon

-0.70

uctive

-0.69

lihood

-0.68

Ïĥ

-0.68

gers

-0.67

ames

-0.67

erd

-0.67

POSITIVE LOGITS

 McMaster

0.90

HAEL

0.84

donald

0.84

 McGill

0.83

awar

0.81

gomery

0.80

ategory

0.77

 Scotia

0.76

ulty

0.74

 Lumpur

0.73

Activations Density 0.008%

names of political figures

occurrences of specific names, particularly individuals and places

No Comments

No Known Activations