INDEX

Explanations

adjectives related to usefulness

oai_token-act-pair · gpt-3.5-turbo

instances of the word "useful."

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

 Fever

-0.73

olina

-0.71

BU

-0.65

opers

-0.64

buck

-0.62

Bom

-0.62

otide

-0.61

uph

-0.60

eree

-0.60

bow

-0.59

POSITIVE LOGITS

 idiots

0.88

tips

0.86

 guiActiveUn

0.79

glers

0.79

fully

0.79

 insights

0.78

 tools

0.77

tip

0.76

 useful

0.74

NESS

0.71

Activations Density 0.027%

adjectives related to usefulness

instances of the word "useful."

No Comments

No Known Activations