INDEX

Explanations

words related to technical elements and instructions, potentially related to a specific field or activity

oai_token-act-pair · gpt-3.5-turbo

the phrase "I'll" in various contexts

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

senal

-0.78

 guiActiveUn

-0.72

cription

-0.71

 à¨

-0.69

template

-0.66

 è£ıç

-0.66

 pane

-0.66

lished

-0.65

ItemImage

-0.61

 smoker

-0.61

POSITIVE LOGITS

uminati

1.29

oyd

1.02

inois

1.02

aday

0.87

ounge

0.86

icit

0.85

owship

0.84

igation

0.83

sburgh

0.82

ateral

0.81

Activations Density 0.021%

words related to technical elements and instructions, potentially related to a specific field or activity

the phrase "I'll" in various contexts

No Comments

No Known Activations