INDEX

Explanations

contractions and words related to past actions

oai_token-act-pair · gpt-3.5-turbo

usage of the verb "to be" in various forms

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

 Approach

-0.68

 Venture

-0.65

*/

-0.63

 Manifest

-0.62

 Testing

-0.62

 Checks

-0.61

 Prototype

-0.59

 Narc

-0.59

 hatch

-0.59

 Campus

-0.58

POSITIVE LOGITS

now

0.85

now

0.83

 hindsight

0.79

 regretted

0.76

 unfocusedRange

0.75

utations

0.74

 remembered

0.74

gered

0.73

ledged

0.70

bol

0.69

Activations Density 0.752%

contractions and words related to past actions

usage of the verb "to be" in various forms

No Comments

No Known Activations