INDEX

Explanations

mentions of specific locations, particularly "Waterloo" and "Wellington"

oai_token-act-pair · gpt-3.5-turbo

references to the location "Waterloo" and its various contexts

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

orb

-0.94

ures

-0.85

âķĲâķĲ

-0.77

orate

-0.75

vana

-0.73

ktop

-0.72

ibilities

-0.72

vg

-0.71

alien

-0.69

android

-0.69

POSITIVE LOGITS

loo

0.98

 Waterloo

0.96

 Mellon

0.95

 Spit

0.92

 Harbour

0.89

town

0.86

 Niagara

0.81

 Royale

0.81

 Peninsula

0.78

 Centre

0.77

Activations Density 0.020%

mentions of specific locations, particularly "Waterloo" and "Wellington"

references to the location "Waterloo" and its various contexts

No Comments

No Known Activations