INDEX

Explanations

numerical thresholds or minimum requirements

oai_token-act-pair · gpt-3.5-turbo

phrases indicating minimum requirements or thresholds

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

 DRAG

-0.76

 Dynamics

-0.73

Reviewer

-0.73

bath

-0.72

ãĥ¼ãĤ¯

-0.64

rend

-0.64

 Brawl

-0.64

selves

-0.63

kai

-0.61

tions

-0.60

POSITIVE LOGITS

uner

0.77

 credible

0.76

 partially

0.73

 toler

0.71

 reputable

0.68

 plausible

0.67

 satisfactory

0.66

egal

0.66

 partly

0.66

 viable

0.66

Activations Density 0.021%

numerical thresholds or minimum requirements

phrases indicating minimum requirements or thresholds

No Comments

No Known Activations