INDEX

Explanations

words related to probability or certainty

oai_token-act-pair · gpt-3.5-turbo

occurrences of specific statistical or mathematical terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 Together

-0.76

MAP

-0.66

Together

-0.66

 Plans

-0.66

founded

-0.63

BAT

-0.61

TAG

-0.60

HUD

-0.60

 Facts

-0.60

 Framework

-0.59

POSITIVE LOGITS

 rarely

1.94

 usually

1.94

 seldom

1.88

 often

1.84

 typically

1.82

 invariably

1.80

often

1.71

usually

1.66

 tended

1.65

 generally

1.62

Activations Density 0.743%

words related to probability or certainty

occurrences of specific statistical or mathematical terms

No Comments

No Known Activations