INDEX

Explanations

locations or place names related to cities, neighborhoods, or venues

oai_token-act-pair · gpt-3.5-turbo

references to geographical locations or neighborhoods

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ischer

-0.86

isc

-0.84

ional

-0.83

INAL

-0.81

vol

-0.79

sonian

-0.79

andum

-0.78

adic

-0.76

inent

-0.75

onymous

-0.73

POSITIVE LOGITS

 Heights

1.22

tones

0.86

dale

0.81

 Village

0.77

ville

0.76

burgh

0.75

peed

0.73

 Traps

0.71

bury

0.71

 Plaza

0.71

Activations Density 0.015%

locations or place names related to cities, neighborhoods, or venues

references to geographical locations or neighborhoods

No Comments

No Known Activations