INDEX

Explanations

phrases structured as ''Like [something]''

oai_token-act-pair · gpt-3.5-turbo

the phrase "Like" used in various contexts, indicating a comparison or expression of similarity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 sight

-0.67

raft

-0.63

 solution

-0.62

 quarters

-0.61

vo

-0.59

arial

-0.58

 restoration

-0.57

 retreat

-0.57

 comm

-0.57

 receiving

-0.56

POSITIVE LOGITS

 Like

3.24

Like

2.47

 LIKE

2.30

 Unlike

1.89

like

1.69

 like

1.56

Unlike

1.49

 Similar

1.44

 Likes

1.43

 Whereas

1.43

Activations Density 0.009%

phrases structured as ''Like [something]''

the phrase "Like" used in various contexts, indicating a comparison or expression of similarity

No Comments

No Known Activations