INDEX

Explanations

quantitative descriptions or assessments

oai_token-act-pair · gpt-3.5-turbo

quantifiable measurements or impressive statistics

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

erent

-0.77

veland

-0.75

theless

-0.69

eeper

-0.68

 Titus

-0.65

tails

-0.65

 stale

-0.65

 wiser

-0.63

eret

-0.63

 inactive

-0.61

POSITIVE LOGITS

ulence

0.78

 Medals

0.74

smanship

0.68

Reviewer

0.67

agos

0.66

 feats

0.65

 Beasts

0.65

 discoveries

0.65

 è£ıç

0.64

 Archdemon

0.64

Activations Density 0.574%

quantitative descriptions or assessments

quantifiable measurements or impressive statistics

No Comments

No Known Activations