INDEX

Explanations

proper names of cities or organizations

oai_token-act-pair · gpt-3.5-turbo

specific names of places or locations

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ipeg

-0.91

ĸļ

-0.87

behind

-0.85

draft

-0.84

sheet

-0.81

wt

-0.78

âĸĪâĸĪâĸĪâĸĪâĸĪâĸĪâĸĪâĸĪ

-0.77

reason

-0.75

drawn

-0.75

know

-0.75

POSITIVE LOGITS

 Parish

1.12

 Pharmaceutical

0.93

 Palace

0.91

 Gardens

0.90

 Springs

0.89

 Technologies

0.89

 Islands

0.88

 Syndrome

0.88

phia

0.86

 Heights

0.84

Activations Density 0.357%

proper names of cities or organizations

specific names of places or locations

No Comments

No Known Activations