INDEX

Explanations

mentions of cycles or repetitive patterns

oai_token-act-pair · gpt-3.5-turbo

references to cycles

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

inez

-0.93

essors

-0.87

orage

-0.83

avez

-0.82

amina

-0.81

imus

-0.76

oaded

-0.76

azeera

-0.75

iciency

-0.74

initions

-0.74

POSITIVE LOGITS

 cycle

1.17

 Cycle

0.99

cycle

0.94

 cycles

0.93

cycles

0.91

erg

0.85

cles

0.84

 repeats

0.77

 Delay

0.73

 alternating

0.73

Activations Density 0.020%

mentions of cycles or repetitive patterns

references to cycles

No Comments

No Known Activations