INDEX

Explanations

terms related to technical topics, particularly related to maps and technology

oai_token-act-pair · gpt-3.5-turbo

references to assumptions and mapping concepts

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

 ÂŃ

-1.00

âĢĳ

-0.88

advertisement

-0.77

 automobile

-0.70

SPONSORED

-0.70

Â®,

-0.68

)—

-0.66

oused

-0.66

"—

-0.66

 Caption

-0.66

POSITIVE LOGITS

lol

1.28

 didnt

1.25

:(

1.23

 doesnt

1.22

 dont

1.21

:)

1.21

lol

1.17

 haha

1.12

 alot

1.11

;)

1.11

Activations Density 0.548%

terms related to technical topics, particularly related to maps and technology

references to assumptions and mapping concepts

No Comments

No Known Activations