INDEX

Explanations

terms indicating something surpassing a limit or threshold

oai_token-act-pair · gpt-3.5-turbo

references to limits or thresholds being surpassed

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

rug

-0.78

choice

-0.68

pring

-0.66

onna

-0.66

Ale

-0.66

seed

-0.65

Downloadha

-0.65

rol

-0.65

roller

-0.64

Sham

-0.64

POSITIVE LOGITS

ingly

0.86

0.82

 expectations

0.81

ceed

0.77

=>

0.76

atos

0.75

0.72

capacity

0.71

 exceed

0.70

imates

0.66

Activations Density 0.017%

terms indicating something surpassing a limit or threshold

references to limits or thresholds being surpassed

No Comments

No Known Activations