INDEX

Explanations

words related to abbreviations, particularly those starting with "ab"

oai_token-act-pair · gpt-3.5-turbo

occurrences of the prefix "ab" in words

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

 Nadu

-0.76

enegger

-0.73

 Maker

-0.69

ancial

-0.69

 Sieg

-0.68

 Savior

-0.68

 Inquisitor

-0.67

 Cups

-0.67

zsche

-0.65

 Grail

-0.65

POSITIVE LOGITS

usable

1.01

raham

1.00

omination

0.99

ject

0.96

ibi

0.96

stract

0.94

obo

0.93

bey

0.92

rog

0.92

duct

0.91

Activations Density 0.012%

words related to abbreviations, particularly those starting with "ab"

occurrences of the prefix "ab" in words

No Comments

No Known Activations