INDEX

Explanations

the word "also"

oai_token-act-pair · gpt-3.5-turbo

the word "also," indicating a focus on additive or supplementary information

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

jam

-0.72

pering

-0.67

Bind

-0.65

agus

-0.62

ysis

-0.62

ASE

-0.61

enter

-0.61

ishers

-0.60

Â£

-0.59

adena

-0.59

POSITIVE LOGITS

 anecd

0.84

 occasionally

0.84

 optionally

0.83

 priv

0.76

 intermitt

0.75

 incorporates

0.74

 noteworthy

0.74

 incidentally

0.73

 includes

0.73

 hinted

0.72

Activations Density 0.109%

the word "also"

the word "also," indicating a focus on additive or supplementary information

No Comments

No Known Activations