INDEX

Explanations

references to a specific word, in this case, "Full"

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "Full"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

emet

-0.72

inations

-0.67

iev

-0.64

hesion

-0.64

ioxide

-0.63

onse

-0.62

trap

-0.62

orks

-0.61

 unsus

-0.60

yrus

-0.59

POSITIVE LOGITS

 Full

3.58

Full

2.47

full

2.13

 FULL

2.11

 full

1.89

 Fuller

1.52

 Fully

1.41

 Partial

1.38

 fuller

1.28

 fullest

1.23

Activations Density 0.009%

references to a specific word, in this case, "Full"

occurrences of the word "Full"

No Comments

No Known Activations