INDEX

Explanations

words related to challenging and difficult situations

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "gru" in various contexts, often associated with challenging or overwhelming experiences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ablishment

-0.74

react

-0.73

Smart

-0.72

Elect

-0.69

Mir

-0.68

onomic

-0.68

imates

-0.68

ropolitan

-0.68

CDC

-0.67

lift

-0.66

POSITIVE LOGITS

gru

1.32

 corrid

0.91

 laure

0.86

 awaited

0.83

 unden

0.83

eling

0.79

Â¯Â¯Â¯Â¯

0.77

fort

0.75

grim

0.75

ache

0.75

Activations Density 0.007%

words related to challenging and difficult situations

occurrences of the word "gru" in various contexts, often associated with challenging or overwhelming experiences

No Comments

No Known Activations