INDEX

Explanations

words related to actions or processes in some sort of structured context

oai_token-act-pair · gpt-3.5-turbo

the presence of action verbs and expressions related to plans or processes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

ONSORED

-0.81

Imagine

-0.75

ModLoader

-0.72

Â¯

-0.69

YL

-0.66

âĢķ

-0.65

SAN

-0.63

 fiasco

-0.63

Whoever

-0.62

Ur

-0.62

POSITIVE LOGITS

 exchanged

0.77

 abound

0.74

 vary

0.70

 everywhere

0.70

 prolifer

0.66

hots

0.66

wana

0.65

 plentiful

0.64

 erupt

0.64

 routinely

0.63

Activations Density 0.503%

words related to actions or processes in some sort of structured context

the presence of action verbs and expressions related to plans or processes

No Comments

No Known Activations