INDEX

Explanations

phrases related to geographic locations

oai_token-act-pair · gpt-3.5-turbo

references to specific names, particularly related to places or brands

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Butterfly

-0.76

forth

-0.75

 Covenant

-0.70

 Ashton

-0.68

WHO

-0.68

 Beyond

-0.67

 Fever

-0.67

 Juliet

-0.66

 Karin

-0.66

 Continued

-0.65

POSITIVE LOGITS

terior

0.99

aiman

0.96

iple

0.93

monary

0.93

umn

0.91

gur

0.90

uxe

0.88

ftime

0.87

uge

0.84

aimon

0.84

Activations Density 0.030%

phrases related to geographic locations

references to specific names, particularly related to places or brands

No Comments

No Known Activations