INDEX

Explanations

entries related to coding instructions or procedures, particularly those involving adding, creating, or updating

oai_token-act-pair · gpt-3.5-turbo

instructions related to programming or technical updates

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

rather

-0.87

albeit

-0.87

Rather

-0.85

entimes

-0.84

ographically

-0.82

iably

-0.82

Perhaps

-0.80

conserv

-0.79

ãģ®éŃĶ

-0.77

Republicans

-0.76

POSITIVE LOGITS

 original

0.97

 clipboard

0.94

 latest

0.89

 destination

0.89

 scanner

0.84

 prev

0.83

 desired

0.82

 english

0.82

 Selected

0.79

 USSR

0.79

Activations Density 0.325%

entries related to coding instructions or procedures, particularly those involving adding, creating, or updating

instructions related to programming or technical updates

No Comments

No Known Activations