INDEX

Explanations

sentences containing phrases expressing intensity, comparison, or evaluation

oai_token-act-pair · gpt-3.5-turbo

predicates and descriptions indicating states of being or existence

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 Deaths

-0.72

ainers

-0.71

 Inventory

-0.69

 Saud

-0.67

 Artists

-0.64

 Factors

-0.64

 Landing

-0.64

 Sources

-0.63

 Hust

-0.61

 Peoples

-0.61

POSITIVE LOGITS

 indistinguishable

1.03

 neither

1.00

 antit

1.00

 reminiscent

1.00

 supposed

0.98

 conducive

0.96

 remotely

0.91

 mutually

0.90

cedented

0.90

 otherwise

0.90

Activations Density 0.205%

sentences containing phrases expressing intensity, comparison, or evaluation

predicates and descriptions indicating states of being or existence

No Comments

No Known Activations