INDEX

Explanations

terms related to geographic features and directions like rivers, tunnels, roads, and pathways

oai_token-act-pair · gpt-3.5-turbo

references to geographical locations and pathways

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Marketable

-0.70

 Evaluation

-0.69

 Judgment

-0.67

 Reign

-0.66

 Competitive

-0.66

rogram

-0.66

ottest

-0.65

urchase

-0.64

 Expression

-0.63

cious

-0.63

POSITIVE LOGITS

 connects

1.41

 connecting

1.31

 linking

1.28

bis

1.25

 travers

1.20

 intersect

1.16

 separates

1.10

 connect

1.08

 routes

1.06

 traverse

1.05

Activations Density 0.266%

terms related to geographic features and directions like rivers, tunnels, roads, and pathways

references to geographical locations and pathways

No Comments

No Known Activations