INDEX

Explanations

phrases expressing preference or suggestion

oai_token-act-pair · gpt-3.5-turbo

expressions of doing something effectively or satisfactorily

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

hyde

-0.82

ategory

-0.72

laus

-0.70

heast

-0.68

adena

-0.68

EStreamFrame

-0.64

anos

-0.63

 furiously

-0.62

mid

-0.61

zanne

-0.60

POSITIVE LOGITS

 behaved

0.76

Initialized

0.71

ector

0.71

NESS

0.69

lied

0.68

ogical

0.68

ECT

0.68

 suited

0.68

iberal

0.66

 tuned

0.65

Activations Density 0.048%

phrases expressing preference or suggestion

expressions of doing something effectively or satisfactorily

No Comments

No Known Activations