INDEX

Explanations

references to specific time periods or durations

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of specific time periods

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

ply

-0.87

ILLE

-0.72

Say

-0.71

CHAT

-0.69

 Cosponsors

-0.69

XY

-0.67

Always

-0.67

PRES

-0.66

SB

-0.66

RAY

-0.66

POSITIVE LOGITS

 periods

1.20

icals

1.08

 period

0.90

 Period

0.81

period

0.80

ional

0.75

ozo

0.74

Ago

0.74

ago

0.74

 spans

0.73

Activations Density 0.021%

references to specific time periods or durations

repeated mentions of specific time periods

No Comments

No Known Activations