INDEX

Explanations

phrases related to specific places or locations

oai_token-act-pair · gpt-3.5-turbo

instances of place names and geographical references

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

uggest

-0.79

 pedals

-0.71

 ingest

-0.71

 efficacy

-0.68

handler

-0.67

iot

-0.67

 Mehran

-0.65

 payload

-0.65

 appre

-0.65

repre

-0.65

POSITIVE LOGITS

 birthplace

0.99

England

0.99

 Ruins

0.98

 suburbs

0.88

opolis

0.87

 borders

0.86

chester

0.84

 England

0.84

 suburb

0.84

 skyline

0.83

Activations Density 0.657%

phrases related to specific places or locations

instances of place names and geographical references

No Comments

No Known Activations