INDEX

Explanations

the word "that" with varying emphasis

oai_token-act-pair · gpt-3.5-turbo

the word "that" used 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 @ 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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 pione

-0.98

ãĤĬ

-0.92

izont

-0.88

 exting

-0.86

ă

-0.85

å§«

-0.85

đ

-0.85

ñ

-0.85

Ĝ

-0.85

ü

-0.84

POSITIVE LOGITS

 there

0.91

 aspect

0.88

 notion

0.88

 they

0.87

 lesson

0.87

 fact

0.87

 phrase

0.81

 possibility

0.80

 distinction

0.80

 idea

0.79

Activations Density 0.154%

the word "that" with varying emphasis

the word "that" used in various contexts

No Comments

No Known Activations