INDEX

Explanations

terms related to rankings and position

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

information related to rankings

oai_token-act-pair · gpt-3.5-turbo

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

ãĥĨãĤ£

-0.85

vous

-0.78

Nou

-0.75

 Spiegel

-0.74

idem

-0.69

 Infinite

-0.67

 Bale

-0.66

igne

-0.66

Syd

-0.66

 Â¯

-0.65

POSITIVE LOGITS

 ranked

1.00

ranked

0.82

Ranked

0.81

inois

0.81

 ranking

0.79

Rank

0.77

hyde

0.74

¶æ

0.72

 bidder

0.70

irection

0.70

Activations Density 0.014%

terms related to rankings and position

information related to rankings

No Comments

No Known Activations