INDEX

Explanations

phrases or words related to delayed or postponed events

oai_token-act-pair · gpt-3.5-turbo

variable-length substrings or prefixes that could represent various forms of ongoing action or condition

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

bots

-0.61

 ethics

-0.58

 squats

-0.58

NESS

-0.58

 manipulation

-0.57

ĺħ

-0.56

 shaming

-0.55

SIZE

-0.55

_-

-0.54

DM

-0.54

POSITIVE LOGITS

ested

0.91

ited

0.89

erest

0.86

ivated

0.84

istant

0.83

aunted

0.82

vered

0.79

oved

0.79

oried

0.79

ained

0.78

Activations Density 0.131%

phrases or words related to delayed or postponed events

variable-length substrings or prefixes that could represent various forms of ongoing action or condition

No Comments

No Known Activations