INDEX

Explanations

key terms related to computer programming and command actions

oai_token-act-pair · gpt-3.5-turbo

specific sequences or substrings within words, particularly those associated with actions or functions

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

 consequential

-0.82

 identification

-0.71

BuyableInstoreAndOnline

-0.68

 inclusive

-0.66

 generic

-0.66

Disclaimer

-0.66

 adolesc

-0.66

 relevance

-0.65

 Byzantine

-0.65

 Meaning

-0.64

POSITIVE LOGITS

wrote

1.25

rolled

1.08

igrated

1.08

uld

1.06

ulates

1.03

ounced

1.00

itates

1.00

ieved

0.98

posed

0.97

elled

0.97

Activations Density 0.234%

key terms related to computer programming and command actions

specific sequences or substrings within words, particularly those associated with actions or functions

No Comments

No Known Activations