INDEX

Explanations

terms related to digital technologies, tools, and platforms

oai_token-act-pair · gpt-3.5-turbo

references to analytics, metrics, and data-related terms

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rolet

-0.53

etheless

-0.52

.''.

-0.47

imated

-0.47

kefeller

-0.46

"},"

-0.46

 endif

-0.45

.}

-0.45

.ï¿½

-0.44

onto

-0.44

POSITIVE LOGITS

we

1.11

 they

0.93

 there

0.88

it

0.84

you

0.81

he

0.76

she

0.75

!,

0.74

+,

0.73

Â®,

0.73

Activations Density 0.715%

terms related to digital technologies, tools, and platforms

references to analytics, metrics, and data-related terms

No Comments

No Known Activations