INDEX

Explanations

instances of the word "though" indicating contrast or concession

oai_token-act-pair · gpt-3.5-turbo

the word "though" in various contexts

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

 nuns

-0.60

ynski

-0.57

inated

-0.57

ploy

-0.56

 roofs

-0.56

ãĥĵ

-0.56

Foot

-0.55

 Fallen

-0.55

let

-0.54

 bark

-0.53

POSITIVE LOGITS

swer

0.75

atcher

0.71

deals

0.70

orm

0.68

heit

0.68

ullah

0.67

ettings

0.66

STATE

0.65

iphany

0.65

tered

0.64

Activations Density 0.022%

instances of the word "though" indicating contrast or concession

the word "though" in various contexts

No Comments

No Known Activations