INDEX

Explanations

instances where a quantity or number of occurrences is mentioned, specifically the phrase "times" followed by a number

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate frequency or repetition

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

Reviewer

-1.07

CHAT

-1.04

actionDate

-0.78

aceous

-0.76

ItemTracker

-0.76

ITAL

-0.76

ARI

-0.72

CVE

-0.71

yet

-0.69

Untitled

-0.69

POSITIVE LOGITS

cale

1.14

manship

0.79

bp

0.78

hops

0.77

bps

0.77

 elapsed

0.76

hower

0.73

times

0.72

 Cups

0.71

mith

0.70

Activations Density 0.025%

instances where a quantity or number of occurrences is mentioned, specifically the phrase "times" followed by a number

phrases that indicate frequency or repetition

No Comments

No Known Activations