INDEX

Explanations

words related to using something as a reference or cue for a particular purpose

oai_token-act-pair · gpt-3.5-turbo

terms that indicate various forms of utility or functionality

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

doms

-0.82

itled

-0.68

qus

-0.68

ymes

-0.63

azes

-0.62

errors

-0.62

Jed

-0.62

lez

-0.62

Os

-0.62

Tuc

-0.61

POSITIVE LOGITS

 tool

0.99

 reminder

0.93

 filler

0.92

 deterrent

0.92

 conduit

0.90

ifier

0.90

 measure

0.89

wark

0.88

 fodder

0.87

 surrogate

0.87

Activations Density 0.240%

words related to using something as a reference or cue for a particular purpose

terms that indicate various forms of utility or functionality

No Comments

No Known Activations