INDEX

Explanations

phrases that indicate a falling or not meeting a standard or expectation

oai_token-act-pair · gpt-3.5-turbo

instances of the phrase "falls short" 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

RAW

-0.70

ellar

-0.63

Tu

-0.61

forces

-0.59

RI

-0.59

CENT

-0.59

itational

-0.58

ellig

-0.58

asi

-0.58

 source

-0.58

POSITIVE LOGITS

eners

0.80

leaf

0.79

xus

0.75

¿½

0.74

akuya

0.74

quartered

0.74

 Parables

0.73

ened

0.72

ening

0.72

sighted

0.72

Activations Density 0.023%

phrases that indicate a falling or not meeting a standard or expectation

instances of the phrase "falls short" in various contexts

No Comments

No Known Activations