INDEX

Explanations

words related to actions and events

oai_token-act-pair · gpt-3.5-turbo

elements related to humor and playful themes

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

ccording

-0.68

ecause

-0.66

Interested

-0.65

hemy

-0.63

rencies

-0.61

 Helpful

-0.59

cale

-0.59

 Marketable

-0.59

ortium

-0.59

ync

-0.59

POSITIVE LOGITS

iest

1.02

 itself

0.91

liest

0.90

osphere

0.86

.}

0.79

 beforehand

0.77

 operator

0.75

ultimate

0.74

 portion

0.71

 thereof

0.69

Activations Density 0.716%

words related to actions and events

elements related to humor and playful themes

No Comments

No Known Activations