INDEX

Explanations

instances where something is being analyzed, assessed, or evaluated

oai_token-act-pair · gpt-3.5-turbo

mentions of the term "review" 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 @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 Hots

-0.71

fall

-0.70

pox

-0.69

bott

-0.69

kt

-0.67

poke

-0.66

brates

-0.63

Corn

-0.63

âĹ¼

-0.62

pots

-0.60

POSITIVE LOGITS

opian

0.99

spective

0.85

 feasibility

0.81

rador

0.78

 judgment

0.78

essee

0.77

 process

0.77

 retrospect

0.75

 evaluations

0.75

eport

0.73

Activations Density 0.033%

instances where something is being analyzed, assessed, or evaluated

mentions of the term "review" in various contexts

No Comments

No Known Activations