INDEX

Explanations

phrases indicating tense, high-stress situations or events

oai_token-act-pair · gpt-3.5-turbo

references to tense situations or conditions that are complex and require careful handling

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

bern

-0.84

ervative

-0.81

onest

-0.78

hirt

-0.75

ervatives

-0.75

leeve

-0.74

tein

-0.73

atron

-0.70

itte

-0.70

pees

-0.69

POSITIVE LOGITS

 unfolding

0.79

 occurring

0.76

 happening

0.75

 turbulence

0.72

 situations

0.72

sit

0.71

 Situation

0.71

 witnessed

0.71

 heightened

0.68

 unfold

0.67

Activations Density 0.551%

phrases indicating tense, high-stress situations or events

references to tense situations or conditions that are complex and require careful handling

No Comments

No Known Activations