INDEX

Explanations

phrases related to top rankings or superlatives

oai_token-act-pair · gpt-3.5-turbo

phrases emphasizing rankings or superlative descriptions

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

ADRA

-0.80

ulhu

-0.75

 Noir

-0.74

 Debor

-0.73

 Dickinson

-0.71

 Richards

-0.70

 Borders

-0.70

 Indigo

-0.70

Cox

-0.69

 Lange

-0.68

POSITIVE LOGITS

sized

1.27

sounding

1.25

generation

1.16

looking

1.15

ever

1.15

exclusive

1.15

ranked

1.12

advertisement

1.08

performing

1.07

ranking

1.05

Activations Density 0.082%

phrases related to top rankings or superlatives

phrases emphasizing rankings or superlative descriptions

No Comments

No Known Activations