INDEX

Explanations

how something is described or explained

oai_token-act-pair · gpt-3.5-turbo

the word "how" in various contexts, indicating descriptions of processes or events

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

odder

-0.73

 Mercenary

-0.65

izu

-0.63

ceptions

-0.63

holder

-0.61

erville

-0.61

actor

-0.61

ature

-0.59

wear

-0.59

ception

-0.59

POSITIVE LOGITS

soever

0.78

beit

0.77

 much

0.69

 pervasive

0.67

-+-+

0.65

ihad

0.64

links

0.64

ells

0.64

ever

0.64

ls

0.63

Activations Density 0.074%

how something is described or explained

the word "how" in various contexts, indicating descriptions of processes or events

No Comments

No Known Activations