INDEX

Explanations

numbers around 200

oai_token-act-pair · gpt-3.5-turbo

references to the number 200 in various contexts

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

warts

-0.93

stal

-0.78

book

-0.77

robe

-0.76

aken

-0.71

Äĩ

-0.70

BOOK

-0.70

cone

-0.70

achi

-0.70

ime

-0.68

POSITIVE LOGITS

0.96

Hz

0.94

mph

0.90

âĸĪ

0.82

MHz

0.79

kHz

0.78

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

0.78

dayName

0.77

MHz

0.77

éĹĺ

0.76

Activations Density 0.021%

numbers around 200

references to the number 200 in various contexts

No Comments

No Known Activations