INDEX

Explanations

key information or specific results within a text

oai_token-act-pair · gpt-3.5-turbo

instances of the word "showed" and its variations

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

amber

-0.60

 indefinite

-0.56

ential

-0.55

 Allies

-0.55

 planetary

-0.54

 term

-0.53

altern

-0.53

 financially

-0.52

 Melody

-0.51

 immediate

-0.51

POSITIVE LOGITS

 showed

3.15

shows

2.05

 demonstrated

1.90

 shows

1.88

 shown

1.78

 exhibited

1.77

 showcased

1.66

 show

1.64

show

1.63

 displayed

1.61

Activations Density 0.015%

key information or specific results within a text

instances of the word "showed" and its variations

No Comments

No Known Activations