INDEX

Explanations

a variety of ranges or options within a context

oai_token-act-pair · gpt-3.5-turbo

instances of the word "range."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

mit

-0.77

der

-0.72

rics

-0.65

rafted

-0.65

hoc

-0.65

Hub

-0.65

äºĶ

-0.65

mond

-0.63

Fe

-0.62

gary

-0.62

POSITIVE LOGITS

 Rover

1.06

finder

1.02

 ranges

1.01

 range

0.91

range

0.88

ahead

0.78

ranging

0.74

 Range

0.72

 ranging

0.66

 lengths

0.66

Activations Density 0.017%

a variety of ranges or options within a context

instances of the word "range."

No Comments

No Known Activations